Unveiling NVIDIA's Video Search Agent running on YOUR Server!

🔥 **NVIDIA NeMo Microservices: Complete Tutorial for Model Fine-Tuning** 🔥 VSS Blueprint: https://nvda.ws/3GVG8wl Read the blog: https://nvda.ws/4j84lgc In this comprehensive tutorial, I'll walk you through setting up NVIDIA NeMo Microservices to implement an efficient data flywheel for your AI projects. Learn how to fine-tune Llama 3.2 1B Instruct model with function calling capabilities using the XLAM Salesforce dataset! ## 📋 What You'll Learn • Complete setup of NeMo Microservices ecosystem • Data processing with NeMo Curator • Model customisation/fine-tuning with NeMo Customizer • Step-by-step guide to add function calling abilities to Llama 3.2 1B ## 🚀 NeMo Microservices Components • NeMo Curator: Data processing • NeMo Customizer: Model fine-tuning • NeMo Evaluator: Model evaluation (3x reduction in APIs) • NeMo Guardrails: Safety compliance (1.4x higher safety) • NeMo Retriever: RAG pipeline • Llama Nimatron: Reasoning LLM ## 💻 Prerequisites • NVIDIA H100 GPUs (used 2x in this tutorial) • NGC API key (generated from the link in the video) • Docker and Minikube • Hugging Face token for dataset access ## ⚙️ Performance Stats • 1.8x faster post-training with NeMo Customizer • 3x reduction in APIs with NeMo Evaluator • 1.4x higher safety compliance with minimal latency using NeMo Guardrails ## 📝 Tutorial Overview 1. Pre-configuration setup (Docker, Minikube, NeMo pods) 2. Data preparation (downloading and formatting XLAM dataset) 3. Fine-tuning Llama 3.2 1B with function calling capabilities 4. Verification and testing of the customised model ## 📚 Resources All code, configuration files, and links mentioned in this tutorial are available in the links below. 💡 First time might seem complex, but subsequent fine-tuning becomes much easier - just change the dataset! If you liked this tutorial, check out my video on "Build a Video Search and Summarisation Agent", where you can: • Process videos on your own server for complete data privacy • Search for key events within videos • Generate AI-powered summaries of video content • Analyse videos for safety compliance and operational efficiency Thanks to NVIDIA for sponsoring this video! ## 🔗 Important Links • Try it yourself: build.nvidia.com • Get NGC API key: ngc.nvidia.com/signin #NVIDIA #NeMo #AI #MachineLearning #LLM #FineTuning #FunctionCalling #Llama #AITutorial #datascience I'll create YouTube timestamps for this video about an NVIDIA AI Blueprint for video search and summarisation. Here's the timestamp format you requested: Timestamp: 0:00 - Introduction 0:26 - NVIDIA AI Blueprint Overview 2:01 - Architecture Explanation 2:39 - User Interaction Process 3:26 - Getting Started with NGC API Key 3:48 - Deploying with Launchables 4:28 - GPU Resource Allocation 4:52 - Docker Services & Components 5:28 - Accessing the User Interface 6:02 - Live Demo: Warehouse Video Analysis 7:22 - Conclusion